Phenotypic adaptation to P-element invasion. A. Diagram of crosses used to produce dysgenic (w¹ × Har) and control hybrids (Har × w¹). Har is a wild type P strain harboring P-elements and w¹ is a laboratory M stain that lacks these transposons. B–D. Egg production (B), hatch rate (C) and fraction of eggs showing wild type dorsal-ventral patterning (D) as a function of hybrid age. Control reciprocal hybrids (blue). Dysgenic hybrids (red). E. DNA damage in hybrid ovaries. Egg chambers were labeled for γ-H2Av, a modified histone associated with DNA breaks, and for DNA. Oocyte nuclei in 2–4 day old reciprocal hybrids do not label for γ-H2Av (Har × w¹ 2–4 day, arrow). Oocyte nuclei in 2–4 day old dysgenic hybrids, by contrast, show prominent γH2Av foci (w¹ × Har 2–4 day, arrow). Oocyte nuclei in 21 day dysgenic hybrids do not label for γH2Av (w¹ × Har 21 day, arrow). Also see Figure S1.

Vasa localization and expression during hybrid dysgenesis. A. At 2–4 days, only low levels of Vasa is present in the germarium of dysgenic hybrids (2–4 Day, w¹ × Har), while Vasa is dispersed in the cytoplasm and concentrated in perinuclear foci (nuage) in reciprocal hybrid controls (2–4 Day, Har × w¹). B. At 21 days, by contrast, Vasa is present in the cytoplasm and nuage in both dysgenic hybrids and reciprocal control. Vasa also shows the expected accumulation at the posterior pole of stage 10 oocytes (right panels). C. Western blot for Vasa in control and dysgenic ovaries. At 2–4 and 21 days, reciprocal hybrids express a species of the expected 72 kD apparent MW. 2–4 day old dysgenic hybrids, by contrast, express low levels of this species and a prominent higher mobility band. At 21 days, however, the 72 kD band is restored. In panels A and B, the distribution of Vasa is shown on the left and a merge of Vasa (green), DNA (Blue) and Lamin-C (red) is on the right. Note that Lamin-C is prominent in the terminal filament cells that occupy the anterior tip of the germarium and in the stalk cells located between egg chambers. Scale bars = 20 μm.

Organization of the piRNA processing machinery. Dysgenic and reciprocal control ovaries were labeled for the PIWI proteins Ago3 (A, D, G), Piwi (B, E, H), and Ago3 (C, F, I). In 2–4 day and 21 day old reciprocal Har × w¹ controls, Aub and Ago3 show the expected localization to nuage, and Piwi is concentrated in nuclei. In 2–4 day old w¹ × Har dysgenic ovaries, by contrast, Aub, Ago3 and Piwi are dispersed (A–C). In 21 day w¹ × Har dysgenic hybrids, by contrast, localization of all three proteins is comparable to reciprocal controls. Pairs of images show the distribution of the indicated PIWI protein on the left and a merge of the PIWI protein (green), DNA (blue), and Lamin-C (red) on the right. Scale bars = 10 μm.

Transposon, gene and piRNA precursor expression. A. P-element transcript levels in dysgenic hybrids (w¹ × Har) and reciprocal controls (Har × w¹). P-element transcript levels were measured by qPCR and are expressed relative to an internal vasa mRNA control. Whole genome tiling array analysis of gene (B) and transposon family (C) expression in 2–4 and 21 day old dysgenic ovaries relative to w¹ controls. Each point represents a single transposon family or protein coding gene. Points above the diagonal are over-expressed in dysgenic hybrids. Points in red indicate significant over-expression (FDR<0.05). R=Correlation coefficient. D. Transcript levels form both strands of a 4^th chromosome piRNA cluster carrying a paternally inherited P-element insertion. RNA was measured by strand specific RT-qPCR using primers that span the P-element insertion site. Transcripts from both strands were significantly elevated in 2–4 day old dysgenic hybrids relative to reciprocal controls. At 21 days, by contrast, dysgenic hybrids and reciprocal controls express similar levels of cluster transcripts. Transcript levels for an unlinked piRNA cluster on chromosome 2L (control) were not altered in the dysgenic hybrids. Bar graphs display mean and standard deviation. Also see Figure S2 and Table S1.

P-element and resident element piRNA expression. piRNA expression was determined by deep sequencing, normalizing for sequencing depth (see Experimental Procedures). A. 2–4 day old dysgenic ovaries (w¹ × Har) express low levels of P-element piRNAs relative to reciprocal controls (C, Har × w¹). At 21 days, by contrast, dysgenic (B) and control (D) ovaries express P-element piRNAs at similar levels. E. Expression of piRNAs matching shared resident elements in 2–4 and 21 day old hybrids, relative to control hybrids. Group III elements, which are enriched in somatic follicle cells, are in red. Group I elements, which are enriched in the germline, are in black. Group II elements show a sense strand bias and are in green. At 2–4 days, dysgenic ovaries over-express group III piRNA and under-express group I piRNAs. This likely reflects developmental arrest prior to germline expansion. At 21 days, by contrast, wild type piRNA expression is restored. Also see Figure S3.

Transposon mobilization during hybrid dysgenesis. A. New transposon insertions were detected by paired-end genome deep sequencing and quantified following normalization to a depth of 18.3 fold. Bars indicate normalized paired end reads associated with new transposon insertions for specific transposon families in 2–4 day old (blue) and 21 day old dysgenic ovaries (red), relative to the parental Har and w¹ genomes. The green bars indicate new insertions in the fertile progeny of 21 day dysgenic females mated to the parental w¹ strain. These insertions are relative to the 21 day old dysgenic females. Most transposon families are active in the dysgenic hybrids, and roo is more active than the P-element trigger for hybrid dysgenesis. All transposon families show reduced activity in the progeny ovaries. B. Site specific insertion and germline transmission of an Ivk element in the 42AB cluster. Shown are the locations of paired end reads defining a new insertion in dysgenic ovaries. Blue bars indicate plus strand reads with their right end mapping to Ivk. Red bars indicate minus strand reads with left ends mapping to Ivk. Grey bars indicate reads that cross the junction between unique genomic sequences and Ivk, which define the insertion site at the nucleotide level. Ivk is not detected in the paternal Har and w¹ strains, but insertion-defining reads are present at 2–4 days, increase at 21 days, and are more abundant in the progeny of 21 day dysgenic hybrids ((w¹ × Har)x w¹. In all three data sets, insertion spanning reads indicate transposition into precisely the same location, which defines novel junction sequences. C. The sequence of plus (blue) and minus (red) strand piRNAs reads mapping to the left hand Ivk junction, with the number of reads for each sequence indicated (reads, left columns). The number of junction specific piRNA reads increases between 2–4 and 21 days, consistent with the increase in insertion matching genome sequencing reads (B). Ivk transposition thus leads to de novo production of junction specific piRNAs. D. Total piRNAs mapping to Ivk also increase between 2–4 and 21 days (Ivk piRNAs). The number of piRNAs that overlap by 10 nt, and the bias toward this overlap, also increase with hybrid age (normalized read pair). Also see Tables S2 and S3.

Genome distribution of inherited transposon insertions. Direct genome sequencing identified 132 transposon insertion sites in 21 day old dysgenic females that were also present in their progeny at 25% or greater penetrance. The genomic distribution of these inherited insertions is indicated, with bar height indicating penetrance in the progeny genome. The 6 insertions that map to piRNA clusters are in red. All of these sites are in pericentromeric heterochromatin or heterochromatin on chromosome 2. Three of these insertions are in the major pericentromeric cluster at 42AB. Also see Tables S4 and S5.